pfsense-mcp CLAUDE.md

pfsense-mcp CLAUDE.md is an instructions file for coding agents from abl030/pfsense-mcp. It costs 2,905 tokens per session, scanned A, original, MIT.

Repository instructions for a server that exposes the pfSense firewall's REST API through MCP, a standard way for AI tools to use external services. The project generates much of its Python code from an OpenAPI API description.

In plain words
What is it for?
Use them when modifying the pfSense MCP server, its code generator, templates, automated tests, or scripts.
Why use it?
They explain which files are authoritative, how to regenerate code safely, and how to test changes against a virtual machine rather than a live system.

Instructions file

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/abl030/pfsense-mcp/claude-md
Clone the repo
git clone --depth 1 https://github.com/abl030/pfsense-mcp

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for pfsense-mcp CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/abl030/pfsense-mcp/claude-md.svg)](https://agentmods.dev/instructions/abl030/pfsense-mcp/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/abl030/pfsense-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/abl030/pfsense-mcp/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 2,905 This file is loaded in full into every session.
When invoked 2,905 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.02905 $0.02905
Opus 5 $0.01452 $0.01452
Sonnet 5 $0.00581 $0.00581
Haiku 4.5 $0.00291 $0.00291

Measured 4d ago against content hash 8596f469d964, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

pfsense-mcp CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

CLAUDE.md · 210 lines

How it starts

The opening of the file, as written. The whole thing — 210 lines — stays where its author put it; the contents beside it link to each section on GitHub.

pfSense MCP Server

Auto-generated MCP server for the pfSense REST API v2. 677 tools generated from openapi-spec.json (OpenAPI 3.0.0, 258 paths, 178 schemas). When pfSense updates their API, pull a new spec and re-run the generator.

Rules

  1. Never manually edit generated/server.py. Fix the generator or templates instead.
  2. openapi-spec.json is the single source of truth. All type information, parameter names, and endpoint structure come from the spec.
  3. Test against the VM, not production. The golden image exists for this purpose.
  4. Expect scripts are fragile but working. Do not change timing, patterns, or shortcuts unless something breaks.
  5. Always use nix develop -c as the default wrapper for repo commands (including python -m generator, pytest, qemu/expect scripts, and any tooling that touches this project).
  6. Testing must be automated. Every feature/fix needs a pytest suite or equivalent that runs without human intervention. No manual-only verification — if it can't be pytest'd, write a script.
  7. Add tests with features. When adding new functionality, always write automated tests in the same change. Tests verify the feature works and prevent regressions.
  8. Sprint progress lives in CLAUDE.md. When work spans multiple sessions, document sprint plans, progress, and outcomes here so future sessions have full context.

Repository Structure

openapi-spec.json            # OpenAPI 3.0.0 spec (input)
generator/                   # Python generator
  __main__.py                # Entry point: python -m generator
  loader.py                  # Load and parse the OpenAPI spec
  naming.py                  # Convert operationIds to tool names
  schema_parser.py           # Extract parameter types from schemas
  context_builder.py         # Build template context for each tool
  codegen.py                 # Render templates and write output
templates/
  server.py.j2               # FastMCP server template
generated/
  server.py                  # The MCP server (677 tools — never hand-edit)
test_modules.py              # Spec-derived pytest suite for module/read-only gating (85 tests)
test_list_tools.py           # Tests for list tool field selection, row filtering, known-fields (19 tests)
bank-tester/                 # AI-driven integration test suite
  run-bank-test.sh           # Orchestrator: boot VM → run tasks → collect results
  generate-tasks.py          # Auto-generate task files from spec + task-config.yaml
  task-config.yaml           # Per-subsystem test values, deps, endpoint types
  analyze-results.py         # Parse results, compute tool coverage
  TESTER-CLAUDE.md           # System prompt for the tester Claude
  tasks/                     # 53 task files
  results/run-*/             # Per-run results
vm/                          # Test infrastructure
  setup.sh                   # Golden image builder (install + firstboot + upgrade)
  install.exp                # Automated pfSense installer
  firstboot.exp              # First-boot configuration
  upgrade-2.8.exp            # pfSense 2.7.2 → 2.8.1 upgrade
research/                    # Analysis documents and MCP best practices

Read the full file on GitHub · 210 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 4d ago First seen · 210 lines · 2,905 tokens per session scan A 8596f469d964

Subscribe to this mod's changes

pfsense-mcp CLAUDE.md is an instructions file published in the GitHub repository abl030/pfsense-mcp (1 stars, last pushed 2mo ago), licensed MIT. It adds 2,905 tokens to every session, about $0.0145 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.